---
title: "Great American Beer Festival Analysis"
author: "Thilini Vithanage, Shradhda Kaki, Sepand Gousheh"
output:
flexdashboard::flex_dashboard:
orientation: columns
social: menu
source_code: embed
---
```{r setup, include=FALSE}
library(flexdashboard)
library(tidyverse)
library(scales)
library(ggspatial)
library(lubridate)
library(sf)
library(viridis)
library(urbnmapr)
library(animation)
library(tidytext)
library(proto)
library(gsubfn)
library(RSQLite)
library(sqldf)
library(gganimate)
library(plotly)
library(maps)
library(tmap)
library(leaflet)
library(tidylo)
library(dbplyr)
library(ggmap)
#loading the data
beer_awards <- readr::read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2020/2020-10-20/beer_awards.csv')
#giving the right order to medals
beer_awards$medal <- factor(beer_awards$medal,
c("Gold", "Silver", "Bronze"),
levels = c("Gold", "Silver", "Bronze"),
ordered = TRUE)
#### Adding a new variable "new_cat" a more general category based on beer categories ####
beer_awards$category <- tolower(beer_awards$category)
gen_cat <- function(x){
ifelse(str_detect(x,'india pale ale'),'India pale ale',
ifelse(str_detect(x,'impe'),'Imperial',
ifelse(str_detect(x,'pale ale'),'Pale ale',
ifelse(str_detect(x,'stout'),'Stouts',
ifelse(str_detect(x,'lager'),'Lager',
ifelse(str_detect(x,'pils'),'Pilsner',
ifelse((str_detect(x,'fruit beer'))|(str_detect(x,'ras')),'Fruit beer',
ifelse(str_detect(x,'hefeweizen'),'Wheat beer',
ifelse(str_detect(x,'witbier'),'Wheat beer',
ifelse(str_detect(x,'weizen'),'Wheat beer',
ifelse((!str_detect(x,'india pale ale') | !str_detect(x,'pale ale'))&
(str_detect(x,'ale')),'Other ale',
ifelse(str_detect(x,'wheat'),'Wheat beer',
ifelse(str_detect(x,'bock'),'Bock',
ifelse(str_detect(x,'porte'),'Porter',
ifelse(str_detect(x,'bitter'),'Bitter','Other'
)))))))))))))))
}
#adding the new category to dataframe
beer_awards <- beer_awards %>% rowwise() %>% mutate(new_cat = gen_cat(category))
#### Filtering for only Texas dataframe ####
texas_beer <- beer_awards %>% filter(state == 'TX')
#finding the top 8 breweries
texas_brewery <- sqldf("select brewery,count(*) as n
from beer_awards
where state = 'TX'
group by brewery
order by n desc
limit 8")
texas_brewery <- unique(texas_brewery$brewery)
#
texas_beer$brewery <- factor(texas_beer$brewery, levels = texas_brewery)
#creating a states sf object for ploting ggplot objects from urbnmapr package
states_sf <- get_urbn_map("states", sf = TRUE)
```
Texas Beer award analysis
=======
Column {.tabset}
-----------------------------------------------------------------------
### City medal comparison
```{r,fig.height= 10,fig.width= 8}
#looking at what cities in Texas won the most medals
plot1 <- texas_beer %>%
group_by(medal,city) %>%
count(medal) %>%
ggplot(aes(x =reorder_within(city,n,medal),y=n,width=0.75)) +
geom_bar(stat = 'identity',aes(fill=medal)) +
scale_fill_manual(values=c('Gold'="#FFCC33", 'Silver'="#999999",'Bronze'= "#CC6600")) +
labs(title = "Great American Beer Festival Medal Winner Analysis (1987-2020)",
subtitle = "Texas Edition",
x ='Brewery city',
y = "Number of Medal Awards",
caption = "source = Great American Beer Festival \nGraph by @wjtorres\nEdited by group 3",
fill = 'Medal') +
theme_minimal() +
theme(
panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
axis.text = element_text(size = 10),
axis.title = element_text(size = 8),
plot.caption = element_text( size = 8),
plot.title = element_text(size = 14),
plot.subtitle = element_text(size = 12),
legend.title = element_text(size = 12),
legend.text = element_text(size = 8),
axis.line.x = element_line(size = 0.5, linetype = "solid", colour = "black"),
axis.line.y = element_line(size = 0.5, linetype = "solid", colour = "black"),
plot.background = element_rect(fill = "#ffffff"),
axis.title.x = element_text(size =12,face='bold'),
axis.title.y = element_text(size =12,face='bold'),
strip.text.y = element_blank()
) +
coord_flip()+
facet_grid(medal~.,scales = "free")+
scale_x_reordered()+
scale_y_continuous(breaks = seq(0,30,by = 5))
plot1
```
### Brewery medal comparison
```{r,fig.height= 10,fig.width= 9}
#brewery medal comparison
plot2 <- texas_beer %>% filter(brewery %in% texas_brewery) %>%
group_by(medal,brewery) %>%
count(medal) %>%
ggplot(aes(x =reorder_within(brewery,n,medal),y=n,width=0.75)) +
geom_bar(stat = 'identity',aes(fill=medal)) +
scale_fill_manual(values=c('Gold'="#FFCC33", 'Silver'="#999999",'Bronze'= "#CC6600")) +
labs(title = "Great American Beer Festival Medal Winner Analysis (1987-2020)",
subtitle = "Texas Edition",
x ='Brewery name',
y = "Number of Medal Awards",
caption = "source = Great American Beer Festival",
fill = 'Medal') +
theme_minimal() +
theme(
panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
axis.text = element_text(size = 10),
axis.title = element_text(size = 8),
plot.caption = element_text( size = 8),
plot.title = element_text(size = 14),
plot.subtitle = element_text(size = 12),
legend.title = element_text(size = 12),
legend.text = element_text(size = 8),
axis.line.x = element_line(size = 0.5, linetype = "solid", colour = "black"),
axis.line.y = element_line(size = 0.5, linetype = "solid", colour = "black"),
plot.background = element_rect(fill = "#ffffff"),
axis.title.x = element_text(size =12,face='bold'),
axis.title.y = element_text(size =12,face='bold'),
strip.text.y = element_blank()) +
coord_flip()+
facet_grid(medal~.,scales = "free")+
scale_x_reordered()+
scale_y_continuous(breaks = seq(0,30,by = 5))
plot2
```
Column
-----------------------------------------------------------------------
### Brewery medals won by beer type
```{r,fig.width=9}
plot4<- texas_beer %>% filter(brewery %in% texas_brewery[1:4]) %>% group_by(brewery,medal,new_cat) %>% count(medal) %>%
ggplot(aes(x=new_cat,y=n))+
coord_flip()+
facet_wrap(~brewery,scales = 'free')+
geom_col(aes(fill=medal),position = position_dodge2(preserve = "single"))+
scale_fill_manual(values=c('Gold'="#FFCC33", 'Silver'="#999999",'Bronze'= "#CC6600"))+
labs(title = "Texas top 4 breweries and awards based on beer type (1987-2020)",
y='Number of awards',x='Beer type',fill='Medals won',
caption = "source = Great American Beer Festival")+
theme_classic()+
theme(strip.text.x = element_text(size = 12,face = 'bold'),
plot.title = element_text(size = 14),
panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
axis.text = element_text(size = 10),
axis.title = element_text(size = 8),
plot.caption = element_text( size = 8),
plot.subtitle = element_text(size = 12),
legend.title = element_text(size = 12),
legend.text = element_text(size = 8),
axis.line.x = element_line(size = 0.5, linetype = "solid", colour = "black"),
axis.line.y = element_line(size = 0.5, linetype = "solid", colour = "black"),
plot.background = element_rect(fill = "#ffffff"),
axis.title.x = element_text(size =12,face='bold'),
axis.title.y = element_text(size =12,face='bold'))
plot4
```
### Top two beer enteries from Texas
```{r,fig.width=9}
plot6.1<- beer_awards %>% filter(new_cat != 'Other' & state == 'TX' &
new_cat %in% c('Lager','Other ale'))%>% group_by(year,new_cat) %>% count(new_cat)%>%
ggplot(aes(x=year,y=n))+geom_bar(stat = 'identity',aes(fill=new_cat))+facet_wrap(~new_cat)+
scale_x_continuous(breaks = seq(1980,2020,by = 5))+
labs(title = "Texas most awarded beer styles (1987-2020)",
y='Number of awards won',x=element_blank(),
caption = "source = Great American Beer Festival")+
theme_classic()+
theme(strip.text.x = element_text(size = 12,face = 'bold'),
plot.title = element_text(size = 14),
legend.position = "none",
panel.spacing = unit(1.5, "lines"),
panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
axis.text = element_text(size = 10),
axis.title = element_text(size = 8),
plot.caption = element_text( size = 8),
plot.subtitle = element_text(size = 12),
legend.title = element_text(size = 12),
legend.text = element_text(size = 8),
axis.line.x = element_line(size = 0.5, linetype = "solid", colour = "black"),
axis.line.y = element_line(size = 0.5, linetype = "solid", colour = "black"),
plot.background = element_rect(fill = "#ffffff"),
axis.title.x = element_text(size =12,face='bold'),
axis.title.y = element_text(size =12,face='bold'))
plot6.1
```
US beer award analysis
===================
```{r,include= FALSE}
us_brewery <- sqldf("select brewery,count(*) as n
from beer_awards
group by brewery
order by n desc
limit 5")
us_brewery <- unique(us_brewery$brewery)
#
us_beer_select <- beer_awards %>% filter(brewery %in% us_brewery)
us_beer_select$brewery <- factor(us_beer_select$brewery, levels = us_brewery)
```
```{r,echo=FALSE,include=FALSE}
states_sf <- get_urbn_map("states", sf = TRUE)
temp <- beer_awards %>% group_by(state,new_cat,medal) %>% count(medal)
temp1 <- inner_join(states_sf,temp,by=c('state_abbv'='state'))
df_plot <- temp1 %>% filter((new_cat %in% c('Lager','Stouts','Pale ale','Pilsner','Wheat beer','India pale ale')))
rm(temp,temp1)
temp <- beer_awards %>% group_by(state,new_cat) %>% count(new_cat)
temp1 <- inner_join(states_sf,temp,by=c('state_abbv'='state'))
df_plot1 <- temp1 %>% filter((new_cat %in% c('Lager','Stouts','Pale ale','Pilsner','Wheat beer','India pale ale')))
rm(temp,temp1)
```
column {.tabset}
------------
### US top 5 brewries with most medals
```{r,fig.height= 10,fig.width= 8}
#brewery medal comparison
plot12 <- us_beer_select %>%
group_by(medal,brewery) %>%
count(medal) %>%
ggplot(aes(x =reorder_within(brewery,n,medal),y=n,width=0.75)) +
geom_bar(stat = 'identity',aes(fill=medal)) +
scale_fill_manual(values=c('Gold'="#FFCC33", 'Silver'="#999999",'Bronze'= "#CC6600")) +
labs(title = "Top 5 US by total medals won (1987-2020)",
subtitle = "Texas Edition",
x ='Brewery name',
y = "Number of Medal Awards",
caption = "source = Great American Beer Festival",
fill = 'Medal') +
theme_minimal() +
theme(
panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
axis.text = element_text(size = 10),
axis.title = element_text(size = 8),
plot.caption = element_text( size = 8),
plot.title = element_text(size = 14),
plot.subtitle = element_text(size = 12),
legend.title = element_text(size = 12),
legend.text = element_text(size = 8),
axis.line.x = element_line(size = 0.5, linetype = "solid", colour = "black"),
axis.line.y = element_line(size = 0.5, linetype = "solid", colour = "black"),
plot.background = element_rect(fill = "#ffffff"),
axis.title.x = element_text(size =12,face='bold'),
axis.title.y = element_text(size =12,face='bold'),
strip.text.y = element_blank()) +
coord_flip()+
facet_grid(medal~.,scales = "free")+
scale_x_reordered()+
scale_y_continuous(breaks = seq(0,30,by = 5))
plot12
```
### Brewery medals won by beer type
```{r,fig.height= 8,fig.width= 10}
plot14<- us_beer_select %>% group_by(brewery,medal,new_cat) %>% count(medal) %>%
ggplot(aes(x=new_cat,y=n))+
coord_flip()+
facet_wrap(~brewery,scales = 'free')+
geom_col(aes(fill=medal),position = position_dodge2(preserve = "single"))+
scale_fill_manual(values=c('Gold'="#FFCC33", 'Silver'="#999999",'Bronze'= "#CC6600"))+
labs(title = "US top 5 breweries and awards based on beer type (1987-2020)",
y='Number of awards',x='Beer type',fill='Medals won',
caption = "source = Great American Beer Festival")+
theme_classic()+
theme(strip.text.x = element_text(size = 12,face = 'bold'),
plot.title = element_text(size = 14),
panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
axis.text = element_text(size = 10),
axis.title = element_text(size = 8),
plot.caption = element_text( size = 8),
plot.subtitle = element_text(size = 12),
legend.title = element_text(size = 12),
legend.text = element_text(size = 8),
axis.line.x = element_line(size = 0.5, linetype = "solid", colour = "black"),
axis.line.y = element_line(size = 0.5, linetype = "solid", colour = "black"),
plot.background = element_rect(fill = "#ffffff"),
axis.title.x = element_text(size =12,face='bold'),
axis.title.y = element_text(size =12,face='bold'))
plot14
```
### beer types
```{r,fig.height= 10,fig.width= 10}
plot5<- beer_awards %>% filter(new_cat != 'Other')%>% group_by(year,new_cat) %>% count(new_cat)%>%
ggplot(aes(x=year,y=n))+geom_bar(stat = 'identity',aes(fill=new_cat))+facet_wrap(~new_cat,scale='free',ncol = 3)+
scale_x_continuous(breaks = seq(1980,2020,by = 5))+
labs(title = "General beer styles based on total number of medals won (1987-2020)",
y='Number of awards',x=element_blank(),
caption = "source = Great American Beer Festival")+
theme_classic()+
theme(strip.text.x = element_text(size = 12,face = 'bold'),
plot.title = element_text(size = 14),
legend.position = "none")
plot5
```
### Interactive US map
```{r}
### load map from library
temp <- beer_awards
temp$STATE<-toupper(temp$state)
temp$state<-NULL
### counting most occurring beers in the state using tidylo library
new_beerlabel <- sqldf("select state,new_cat,count(*)as n
from beer_awards
where new_cat != 'Other'
group by state,new_cat")
new_beerlabel %>% group_by(state) %>%
mutate(my_ranks = order(order(n, decreasing=TRUE))) -> new_beerlabel
new_beerlabel <- new_beerlabel %>% filter(my_ranks == 1) %>% select(state,new_cat)
colnames(new_beerlabel) <- c('STATE','beer')
# count of total medals state wise
total_medal <- sqldf("select state as STATE,count(*) as n
from beer_awards
group by state")
# adding map elements
Us_map<-get_urbn_map("states",sf=TRUE)
new_map<-inner_join(Us_map,new_beerlabel,by=c('state_abbv'='STATE'))
new_map<-inner_join(new_map,total_medal,by=c('state_abbv'='STATE'))
addresses<-unique(new_map$address)
new_map$geometry<-NULL
states <- geojsonio::geojson_read("gz_2010_us_040_00_500k.json", what = "sp")
colnames(new_map)[3]="NAME"
s<-merge(x=states,y=new_map,by='NAME',all.x=TRUE)
#Create a final spatialPolygonsDataframe to create map
require(sp)
final_table<-merge(states,s,by="NAME")
###making map interactive using leaflet
bins<-c(0,10,20,50,100,500,1000) #Create a color palette with Bins
pal<-colorBin("BuPu" ,domain=final_table,bins =bins)
labels <- paste( #Create a label with state name and their number one famous beer
final_table@data$NAME,"
",
final_table@data$beer
) %>%
lapply(htmltools::HTML)
m<-leaflet(final_table)%>%
setView(-96, 37.8, 4)%>%
addProviderTiles(providers$Stamen.Toner)%>%
addPolygons( data=final_table,
weight=1,
color="#660000",
smoothFactor=1,
dashArray=3,
fillOpacity=0.7)
m <- m %>% addPolygons(
fillColor = ~pal(n),
weight = 2,
opacity = 1,
color = "black",
dashArray = "3",
fillOpacity = 0.7,
highlight = highlightOptions(
weight = 5,
color = "#660000",
dashArray = "",
fillOpacity = 0.7,
bringToFront = TRUE),
label = labels,
labelOptions = labelOptions(
style = list("font-weight" = "normal", padding = "3px 8px"),
textsize = "15px",
direction = "auto"))
m %>% addLegend(pal = pal, values = ~n, opacity = 0.7, title = "Total medal and their famous beer",
position = "bottomright")
```
column
--------------------
### beer types
```{r,fig.width= 9}
beer_type <- sqldf("select new_cat,count(*) as n
from beer_awards
where new_cat != 'Other'
group by new_cat
order by n desc
limit 5")
beer_type <- unique(beer_type$new_cat)
plot3 <- beer_awards %>% filter(new_cat %in% beer_type) %>%
group_by(new_cat) %>%
count(new_cat) %>%
ggplot(aes(x =reorder(new_cat,n),y=n,width=0.75)) +
geom_bar(stat = 'identity',fill='#42c20c') +
labs(title = "Most awarded beer types (1987-2020)",
x ='Beer Type',
y = "Number of Medal Awards",
caption = "source = Great American Beer Festival",
fill = 'Medal') +
theme_minimal() +
theme(
panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
axis.text = element_text(size = 10),
axis.title = element_text(size = 8),
plot.caption = element_text( size = 8),
plot.title = element_text(size = 14),
plot.subtitle = element_text(size = 12),
legend.title = element_text(size = 12),
legend.text = element_text(size = 8),
axis.line.x = element_line(size = 0.5, linetype = "solid", colour = "black"),
axis.line.y = element_line(size = 0.5, linetype = "solid", colour = "black"),
plot.background = element_rect(fill = "#ffffff"),
axis.title.x = element_text(size =12,face='bold'),
axis.title.y = element_text(size =12,face='bold')
#strip.text.y = element_blank(), #I left this in because of color blind people
) +
coord_flip()
plot3
```
### map
```{r,fig.height= 8,fig.width= 8, fig.fullwidth=TRUE}
plot10<- ggplot()+ theme_bw()+
geom_sf(data = states_sf,fill='#ffffff',size = 0.05)+
geom_sf(data = df_plot1,aes(fill=n))+
geom_sf_text(data = df_plot1,aes(label=state_abbv),size=2)+
scale_fill_gradient(low = "#e5ffcc", high = "#3d700b", na.value = NA)+
facet_wrap(~new_cat,ncol = 2)+ coord_sf(datum = NA)+
labs(title = 'Total number of Medals won by breweries in each state (1987-2020)\nbased on beer types',
x=element_blank(),y=element_blank(),fill='Medals won\nscale')+
theme(strip.text.x = element_text(size = 12,face = 'bold'))
plot10
```
